Re: atof problem

[boldt@xxxxxxxxxx]

Tim wrote:
>This may have been answered beofre, but I just wanted to verify what I did
>is ok, or if I should have done it a different way.
>
>I am reading a comma delimited file, and one of the fields is a dollar
>amount.  I want to take the dollar amount which is now in a text string of
>10 character or so, ie: '123.45'
>
>and I want to convert it to a zoned 17,2 but when I used the atof
operation,
>I was always get a value 1 penny less that the string that I initially
read.
>
>So  I added the (H) to the eval command and it solved the problem..
>c                   eval(h)   CheckAmount  = atof(fld8)
>
>My question is this, is that OK?  Or should I get the dollar amount from
the
>string a different way?
>Also I did add the H spec to respect decimal position, but that didn't
help
>the problem...


This is working pretty much as expected.  Floating point is
generally used for modelling physical processes involving
measured values that are typically accurate to only so many
significant digits.  As a result, floating point values are
inherantly inaccurate.  Most integral values, such as
monetary amounts cannot be represented with 100% accurate
precision.

Regarding using half-adjust, that should be fine provided
the precision of your result is not too great.  Before
putting your code into production, make sure you feed it the
maximum possible numeric value into the character field and
see what happens.  I think you'll find that the maximum
number of digits you can safely work with is about 15.  Any
more than that and you risk either losing precision in the
least significant digits or numeric overflow.

Regarding the H-Spec option EXPROPTS(*RESDECPOS), that only
has an effect for decimal expressions.  Since you're dealing
with a float value, it has no effect here.

I'd recommend using a different technique to convert char
strings into decimal values.  Although atof and half-adjust
may work with reasonably small values, it doesn't scale
higher than around 15 digits.  Now you may well be aware of
that limitation, but the next programmer who works on your
code may not be.  At the very least, document the limits
clearly when using the atof technique.

Cheers!  Hans

Hans Boldt, ILE RPG Development, IBM Toronto Lab, boldt@ca.ibm.com


+---
| This is the RPG/400 Mailing List!
| To submit a new message, send your mail to RPG400-L@midrange.com.
| To subscribe to this list send email to RPG400-L-SUB@midrange.com.
| To unsubscribe from this list send email to RPG400-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---